PYSEC-2026-3133 TensorFlow vulnerable to `CHECK` failures in `FractionalAvgPoolGrad`

Impact The implementation of FractionalAvgPoolGrad does not fully validate the input originputtensorshape. This results in an overflow that results in a CHECK failure which can be used to trigger a denial of service attack. python import tensorflow as tf overlapping = True originputtensorshape =...

PYSEC-2026-3243 TensorFlow vulnerable to `CHECK` failure in `TensorListReserve` via missing validation

Impact In core/kernels/listkernels.cc's TensorListReserve, numelements is assumed to be a tensor of size 1. When a numelements of more than 1 element is provided, then tf.rawops.TensorListReserve fails the CHECKEQ in CheckIsAlignedAndSingleElement. python import tensorflow as tf...

PYSEC-2026-3237 TensorFlow vulnerable to integer overflow in math ops

Impact When RangeSize receives values that do not fit into an int64t, it crashes. cpp auto size = std::isintegral::value ? Eigen::numext::abslimit - start + Eigen::numext::absdelta - T1 / Eigen::numext::absdelta : Eigen::numext::ceil Eigen::numext::abslimit - start / delta; // This check does not...

TensorFlow vulnerable to `CHECK` fail in `RaggedTensorToVariant`

ImpactIf RaggedTensorToVariant is given a rtnestedsplits list that contains tensors of ranks other than one, it results in a CHECK fail that can be used to trigger a denial of service attack.pythonimport tensorflow as tfbatchedinput = Truertnestedsplits = tf.constant0,32,64, shape=3,...

TensorFlow vulnerable to `CHECK` fail in `RaggedTensorToVariant`

ImpactIf RaggedTensorToVariant is given a rtnestedsplits list that contains tensors of ranks other than one, it results in a CHECK fail that can be used to trigger a denial of service attack.pythonimport tensorflow as tfbatchedinput = Truertnestedsplits = tf.constant0,32,64, shape=3,...

PYSEC-2026-3212 TensorFlow vulnerable to `CHECK` fail in `RaggedTensorToVariant`

Impact If RaggedTensorToVariant is given a rtnestedsplits list that contains tensors of ranks other than one, it results in a CHECK fail that can be used to trigger a denial of service attack. python import tensorflow as tf batchedinput = True rtnestedsplits = tf.constant0,32,64, shape=3,...

PYSEC-2026-3348 TensorFlow vulnerable to `CHECK` fail in `RaggedTensorToVariant`

Impact If RaggedTensorToVariant is given a rtnestedsplits list that contains tensors of ranks other than one, it results in a CHECK fail that can be used to trigger a denial of service attack. python import tensorflow as tf batchedinput = True rtnestedsplits = tf.constant0,32,64, shape=3,...

PT-2026-59982

Impact An input encoded that is not a valid CompositeTensorVariant tensor will trigger a segfault in tf.raw ops.CompositeTensorVariantToComponents. python import tensorflow as tf encode = tf.raw ops.EmptyTensorListelement dtype=tf.int32, element shape=10, 15, max num elements=2 meta= ""...

PT-2026-59921

Impact When TensorListFromTensor receives an element shape of a rank greater than one, it gives a CHECK fail that can trigger a denial of service attack. python import tensorflow as tf arg 0=tf.random.uniformshape=6, 6, 2, dtype=tf.bfloat16, maxval=None arg 1=tf.random.uniformshape=6, 9, 1, 3,...

PT-2026-59807

Impact array ops.upper bound causes a segfault when not given a rank 2 tensor. Patches We have patched the issue in GitHub commit 915884fdf5df34aaedd00fc6ace33a2cfdefa586. The fix will be included in TensorFlow 2.13. We will also cherrypick this commit in TensorFlow 2.12.1. For more information...

PT-2026-60015

Summary Integer truncation of tensor dimensions in vLLM's GGUF dequantize kernels csrc/quantization/gguf/gguf kernel.cu causes partial tensor processing. The output tensor is allocated at full size via torch::empty uninitialized memory, but the dequantize CUDA kernel processes only a truncated...

PT-2026-59854

Impact An input encoded that is not a valid CompositeTensorVariant tensor will trigger a segfault in tf.raw ops.CompositeTensorVariantToComponents. python import tensorflow as tf encode = tf.raw ops.EmptyTensorListelement dtype=tf.int32, element shape=10, 15, max num elements=2 meta= ""...

PT-2026-59945

Impact array ops.upper bound causes a segfault when not given a rank 2 tensor. Patches We have patched the issue in GitHub commit 915884fdf5df34aaedd00fc6ace33a2cfdefa586. The fix will be included in TensorFlow 2.13. We will also cherrypick this commit in TensorFlow 2.12.1. For more information...

PT-2026-59910

Impact The implementation of FractionalAvgPoolGrad does not fully validate the input orig input tensor shape. This results in an overflow that results in a CHECK failure which can be used to trigger a denial of service attack. python import tensorflow as tf overlapping = True orig input tensor...

PT-2026-59121

There exists a use after free vulnerability in Reverb. Reverb supports the VARIANT datatype, which is supposed to represent an arbitrary object in C++. When a tensor proto of type VARIANT is unpacked, memory is first allocated to store the entire tensor, and a ctor is called on each instance...

PT-2026-59722

Impact If QuantizedAvgPool is given min input or max input tensors of a nonzero rank, it results in a segfault that can be used to trigger a denial of service attack. python import tensorflow as tf ksize = 1, 2, 2, 1 strides = 1, 2, 2, 1 padding = "SAME" input = tf.constant1, shape=1,4,4,2,...

PT-2026-59730

Impact Constructing a tflite model with a paramater filter input channel of less than 1 gives a FPE. Patches We have patched the issue in GitHub commit 34f8368c535253f5c9cb3a303297743b62442aaa. The fix will be included in TensorFlow 2.12. We will also cherrypick this commit on TensorFlow 2.11.1...

PT-2026-59906

Impact When mlir::tfg::TFOp::nameAttr receives null type list attributes, it crashes. cpp StatusOr GraphDefImporter::ArgNumTypeconst NamedAttrList &attrs, const OpDef::ArgDef &arg def, SmallVectorImpl &types // Check whether a type list attribute is specified. if !arg def.type list attr.empty if...

PT-2026-59839

Impact When printing a tensor, we get it's data as a const char array since that's the underlying storage and then we typecast it to the element type. However, conversions from char to bool are undefined if the char is not 0 or 1, so sanitizers/fuzzers will crash. Patches We have patched the issu...

PT-2026-59932

Impact The implementation of BlockLSTMGradV2 does not fully validate its inputs. - wci, wcf, wco, b must be rank 1 - w, cs prev, h prev must be rank 2 - x must be rank 3 This results in a a segfault that can be used to trigger a denial of service attack. python import tensorflow as tf use peephol...